W3C home > Mailing lists > Public > www-validator@w3.org > October 2000

Re: User-Agent choice suggestion

From: Konstantin Riabitsev <graf@relhum.org>
Date: Fri, 27 Oct 2000 20:57:39 -0400 (EDT)
To: Terje Bless <link@tss.no>
Cc: www-validator@w3.org
Message-ID: <20001028065422.C1563@graf.etel.ru>
On Sat, Oct 28, 2000 at 12:58:18AM +0200, Terje Bless wrote:

> >This is "broken by design". You will have problems [...]
> Yes, but I think perhaps emphasis should be on __by_design__ rather then
> __broken__ here. It's broken not so much inherently, but rather because it
> will impose a significant burden on the developers to keep the different
> versions in sync and degrading gracefully. However, User-Agent sniffing to
> deliver the most appropriate content to the User-Agent in question can, if
> implemented and maintained well, _improve_ accessability for the site.

That's exactly the reason why I did it. I am aware of possible problems
with cache managers, but every page I give is in fact one HTML/XHTML 
standard or the other. It's mostly a matter of getting around buggy
browsers while delivering the thinest design, because in this part of the
world speed DOES matter a lot.

I am well-aware that one can have a decent design that will look good on
all platforms. However, I, for one, like dark backgrounds and it's either
writing everything in "Transitional" where bgcolor is allowed, or cringing
when seeing my site in a non-CSS browser. I also like to have my menus by
the side of the page text, but I can't use layered positioning or other
advanced CSS stuff if I want people with Netscape-3 and Netscape-4 with
its questionable support for layers to see it.

I think User-Agent "sniffing" can definitely be a useful tool. This way I
can give recent browsers the thinest and the meanest XHTML+CSS, while
folks with older browsers can still see my site without its design being

> Could someone who frequents the JavaScript sites contribute common
> User-Agent strings (along with their associated common name (e.g. "MSIE
> 4.01") so I have something to work with? Konstantin, what's your setup
> like?

As I said, this doesn't have to be on the main page of validator.w3.org,
maybe on some "advanced settings" page. I believe I've seen a file with
User-Agent strings on www.browser.org, but I may be mistaken.

My setup -- since I use Cocoon to do the browser sniffing, it's almost
exactly from their preferences file. I think that having only major
versions of each browser is sufficient. Like, MSIE-3, MSIE-4, MSIE-5,
Netscape-3, Netscape-4, Mozilla5/Netscape6, Opera-3, Opera-4, Lynx, Amaya.
I think that would be sufficient for >95% of people. For the remaining 5%
there can be an "other" option with a text-input field -- if they have
written something for an exotic browser, they most likely can provide the
User-Agent line for it. ;)

Several Unix tools have this ability -- that's how I check my site right
now. I run wget with a --user-agent=STRING option and pipe it to tidy and
that does what I need. However, users on other platforms might not have
this available.

 0>  Konstantin Riabitsev (Mr. Icon) -> icon@mricon.com
/ )  My XML-XSLT powered website -> http://www.mricon.com/
 ~   XMMS playing -> Eagles  -  Hotel California
Received on Friday, 27 October 2000 21:05:02 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 14:17:28 UTC